Feasibility of Home-Based Neurologic Music Therapy for Behavioral and Psychological Symptoms of Dementia: A Pilot Study

Abstract

Family caregivers often feel ill-equipped to handle bothersome behavioral and psychological symptoms of dementia, such as agitation, apathy, and sleep disturbances, leading to increased caregiver distress and nursing home placement for people with dementia. Therapies for such symptoms are currently limited and non-pharmacological options are preferred, given potential side effects of medications. Neurologic music therapy (NMT) could provide an additional treatment option for managing behavioral and psychological symptoms for community-dwelling people with dementia and their caregivers. This pilot study sought to evaluate the feasibility, acceptability, and effectiveness of home-based NMT for behavioral and psychological symptoms of dementia. Eighteen persons with dementia-caregiver dyads were enrolled to receive one-hour weekly sessions of home-based NMT for 6 weeks. Demographic, quality of life, neuropsychiatric symptom, and caregiver burden and self-efficacy information was collected at baseline, 6 weeks, and 12 weeks. Seven dyads (38.9%) withdrew from therapy before completing all sessions; these participants had higher Neuropsychiatric Inventory scores and were of older age at baseline. For those who completed therapy, neuropsychiatric symptom scores improved at 6 weeks, an effect that was sustained at 12 weeks. No other outcome measures changed significantly after therapy. Initiating NMT too late in the course of dementia, when behavioral symptoms are already present, may be impractical for people with dementia and increase caregiver stress, even when provided within the home. Introducing and incorporating the principles of NMT earlier in the course of dementia could allow for increased comfort and benefit for people with dementia and their caregivers.

Dementia is an increasingly common clinical disorder characterized by significant cognitive dysfunction and associated functional impairment, currently affecting nearly 50 million people worldwide (Prince et al., 2015). Over 70% of people with dementia (PWD) reside in the community and rely heavily upon family and informal caregivers (Black et al., 2013). These caregivers face many challenges, chief among these being the behavioral and psychological symptoms of dementia (BPSD), which affect up to 90% of PWD (Nogales-Gonzalez, Romero-Moreno, Losada, Marquez-Gonzalez, & Zarit, 2015; Staedtler & Nunez, 2015). BPSD include agitation, sun-downing (worsening confusion in the early evening), wandering, sleep disturbances, and apathy. Family caregivers often do not feel equipped to handle behavioral symptoms, leading to high levels of caregiver burnout and increased rates of nursing home placement for PWD (Jennings et al., 2015). Current therapies for BPSD are limited and non-pharmacologic interventions are preferred (Gitlin, Kales, & Lyketsos, 2012), given the limited efficacy and high side effect risk profile of psychotropic medications in PWD (Schneider, Dagerman, & Insel, 2006; Tan et al., 2015; Wang et al., 2005). Yet, there is little specific guidance about exactly which non-pharmacological methods are efficacious for BPSD, especially for those that would be appropriate and feasible for use by caregivers in the home setting.

To date, behavioral management techniques and music therapy have shown the most consistency across studies for effectiveness in treatment of BPSD, but evidence to support their use is mostly based on small studies performed in nursing home settings (Azermai et al., 2012). A recent Cochrane Review investigated the effectiveness of music-based therapeutic interventions for a variety of symptoms in dementia, including quality of life, mood, agitation, and cognition (van der Steen et al., 2018). This meta-analysis included 21 studies, with a total of 890 participants, all of which were performed within institutional settings rather than the home. While there was some evidence that at least five sessions of broadly defined music interventions could improve depression and overall behavioral symptoms for nursing home residents with dementia, there were weaker results for quality of life, anxiety, cognition, or agitation. Furthermore, very few studies included any outcomes measured more than 4 weeks after the intervention, making it difficult to support long-lasting effects.

While a number of studies have been performed using music interventions for dementia symptoms, fewer have investigated neurologic music therapy (NMT) specifically. NMT involves the therapeutic application of music to cognitive, sensory, and motor symptoms of neurological diseases, harnessing the potential functional changes that music may impart to the brain (Thaut & Hoemberg, 2014). Taking a scientific approach to address a specific brain function, NMT employs standardized techniques that are adaptable to an individual’s neurological symptoms and needs. The principles of NMT dovetail well with recommended approaches for the management of BPSD; the DICE (Describe, Investigate, Create, Evaluate) method (Kales, Gitlin, & Lyketsos, 2014) provides a framework of behavioral and environmental modification strategies in dementia. The DICE method allows for personalization of treatment to the individual with dementia and their particular symptoms, providing guidance to caregivers and clinicians to develop a management plan. NMT is well aligned with this framework and could offer a new method of non-pharmacological treatment of dementia symptoms that affect quality of life, emotional well-being, and likelihood of nursing home placement. Additionally, bringing NMT into the home setting and educating family caregivers about its techniques, rather than waiting for institutionalization of PWD, could not only improve BPSD, but also decrease caregiver burden and therefore nursing home placement. Home-based NMT would also relieve the stress of traveling outside the home to an external clinic or office to receive such therapies.

Certain NMT techniques hold particular promise in addressing BPSD, by intentionally focusing on cognitive and behavioral neurological functions (Thaut, 2005; Thaut & Hoemberg, 2014). Musical Sensory Orientation Training (MSOT) involves the presentation of music to stimulate arousal and facilitate meaningful responsiveness and orientation to time, place, and person. Musical Attention Control Training (MACT) uses musical elements to cue different musical responses in order to practice sustained, selective, divided, and alternating attention functions. Associative Mood and Memory Training (AMMT) utilizes music to alter mood and enhance mood-dependent memories that can enhance recall of positive memories. Music Psychosocial Counseling (MPC) uses musical performance to address issues of mood control, affective expression, cognitive coherence, reality orientation, and appropriate social interaction to facilitate psychosocial function. In order to tailor an NMT protocol towards BPSD, features of the above therapy techniques could be combined to address multiple symptoms.

Based on the recent literature and gaps in the field, there is a pressing need for safe and effective techniques for the management of BPSD, especially within the home setting, that have durable effects. It is possible that home-based NMT could fill this gap for PWD and their caregivers who reside in the community, rather than in an institutional setting. The primary goal of this pilot study was to determine if home-based NMT was feasible for PWD and their caregivers. As exploratory outcomes, the potential benefit of NMT on frequency and severity of neuropsychiatric symptoms, quality of life, functionality, and caregiver distress and self-efficacy was also evaluated.

Methods

Participants

Participants included 18 PWD, as defined as clinically significant impairment in two cognitive domains and associated impairment in functional abilities, and their identified caregivers. This sample size was primarily determined by funding available for the costs of music therapy sessions for this pilot study. Participants were recruited from the Memory Disorders Clinic, Movement Disorders Clinic, and Neuropalliative and Supportive Care Clinic at the University of Colorado Hospital. Individuals with dementia of any neurological cause, as diagnosed by a neurologist with sub-specialty training in either behavioral neurology or movement disorders at our institution, based on clinical, neuropsychological, and neuroimaging results, as available in the medical record, were included. Only participants with behavioral (e.g., agitation, aggression, wandering, repeating questions) or psychological (e.g., apathy, hopelessness, loneliness, anxiety, hallucinations) symptoms related to their dementia were included. While BPSD can occur throughout the course of dementia, including at the mild cognitive impairment stage (Monastero, Mangialasche, Camarda, Ercolani, & Camarda, 2009), they are often more prominent in the later stages of the disease course (Cerejeira, Lagarto, & Mukaetova-Ladinska, 2012). Therefore, while included participants were expected to skew toward the moderate-to-severe end of the dementia spectrum, dementia of any severity stage was accepted.

Participants were required to be living at home and have a dedicated caregiver in the home who was also willing to participate. Participants were excluded if they were deaf or severely hearing impaired or if they did not speak English. The study was approved by the Colorado Multiple Institutional Review Board and all participants, including PWD and their caregivers, underwent an informed written or oral consent process at the initial screening. For PWD who lacked capacity to consent, based on screening questions by a behavioral neurologist (S.K.H), they provided informed oral assent with formal caregiver consent.

Procedures

Study visits and music therapy sessions took place during a 14-month period between March 2016 and May 2017. Data were collected through self-reported questionnaires, for both PWD and caregivers, and clinician-administered standardized assessments. PWD and caregivers who were unable to travel to the University of Colorado Hospital campus for in-person study visits were provided the option of completing interviews and questionnaires over the phone with research personnel. Demographic and medical information, including etiology of dementia and duration of dementia symptoms, was collected by interview. Baseline scores on the Montreal Cognitive Assessment (MoCA) (Nasreddine et al., 2005) were obtained from medical record review.

A standard battery of validated scales was administered to participants (Table 1), evaluating baseline levels of quality of life (Quality of Life in Alzheimer’s Disease scale [QOL-AD] [Thorgrimsen et al., 2003]); functional disability (Disability Assessment for Dementia [DAD] [Gelinas, Gauthier, McIntyre, & Gauthier, 1999]); neuropsychiatric symptoms (Neuropsychiatric Inventory [NPI] Caregiver Interview [Cummings et al., 1994]); caregiver burden (Zarit Burden Interview [ZBI] [Zarit, 2008]); and caregiving self-efficacy (Revised Scale for Caregiving Self-Efficacy [RSCSE] [Steffen, McKibbin, Zeiss, Gallagher-Thompson, & Bandura, 2002]). Self-efficacy is defined as one’s belief in one’s ability to succeed in a specific situation or to accomplish a task (Bandura, 1994); caregiving self-efficacy is defined as a caregiver’s confidence in their abilities to adequately complete caregiving tasks, including obtaining respite, controlling upsetting thoughts about caregiving, and responding to disruptive patient behaviors (Steffen et al., 2002). Outcome measures were administered at the 1) initial screening visit (baseline); 2) after completion of the home-based NMT protocol (6-week visit); and 3) 6 weeks after completion of the home-based NMT protocol (12-week visit) as outlined in Table 1.

Music Therapy Protocol

The intervention protocol for this pilot was designed by a board-certified music therapist (MT-BC) with additional training in NMT techniques (R.S.), and was reviewed and implemented by a total of four MT-BCs who also have training in NMT intervention techniques. The goal of this specific NMT protocol for BPSD was to potentially enhance cognitive function more directly, including attention, executive function, and memory, but also to provide a focused activity that could be performed together by a PWD and their caregiver, in order to redirect bothersome behaviors and promote a sense of collaboration between them. Fostering non-verbal communication through NMT could also serve to reduce BPSD (Dassa, 2018) by allowing PWD to communicate their needs without resorting to distressing behaviors.

A combination of MSOT, MACT, and AMMT techniques were incorporated into our home-based NMT protocol for BPSD, based on indications from extensive review of both the music therapy and music neuroscience literature, with specific evidence-based features outlined below. This study represents the first utilization of this developed NMT protocol. The protocol was slightly modified for PWD with 1) mild-to-moderate stage dementia or 2) moderate-to-severe dementia, based on staging by a behavioral neurologist (S.K.H.) after interview with the PWD and caregiver (Supplementary Data 1). Slight changes to tasks were made to accommodate for potential limitations in participation for those at the moderate-to-severe stage of dementia, as outlined below.

All interventionists were trained by the MT-BC (R.S.) who designed the protocol to ensure standardized and consistent protocol implementation. Sessions were carried out in the PWD’s home on a weekly basis for a total of 6 weeks. The first therapy session lasted a total of 90 min, to allow for assessment of the participant and familiarization with the therapy; the remaining five sessions were 60 min in length. The caregiver was required to be present for all sessions in the 6-week treatment period, in order to provide the caregiver their own education regarding NMT techniques, as well as to improve positive social and emotional interactions within the session between the PWD and caregivers (Clair & Ebberts, 1997; Hammar, Emami, Gotell, & Engstrom, 2011; Pollack & Namazi, 1992).

The instruments and materials utilized in each session were as follows: one guitar, a basket with a variety of small percussion instruments, one Easycussion pentatonic xylophone, small dry erase board with markers, song list, theme-based visual and tactile aids, theme-based music playlists on a tablet for recorded music listening, and caregiver education sheets. The sessions began with MSOT techniques, including both a sing-along and a play-along portion with an instrument of their choice in order to immediately increase arousal and alertness responses (Clair, 1996; Van de Winckel, Feys, De Weerdt, & Dom, 2004). The sessions gradually increased in complexity, moving from basic arousal and sensory orientation using MSOT into higher level attention tasks utilizing MACT. Music-based tasks focused on sustained and selective attention were included due to some evidence demonstrating improvements on standardized attentional tests following high-frequency music attention-based sessions (Ceccato et al., 2012). Specific exercises included a simple sustained attention and executive function task, in the form of xylophone improvisation for participants with mild-moderate dementia, and a simple sustained attention task using a rain stick, which required the participant to flip it over to re-start the sound on cue, for participants with moderate-severe dementia. Fill-in-the blank singing tasks, designed to address both sustained attention and lyric memory, were also performed (Ceccato et al., 2012; Cuddy et al., 2012), as were more complex sustained attention tasks using color-coded bells.

The last portion of the NMT protocol was music and memory reminiscence through AMMT. Previous research has indicated that not only is musical memory (i.e., memory of lyrics and melody) often maintained throughout the course of dementia (Cuddy & Duffin, 2005), but that autobiographical memories associated with both strong emotion and music can be more easily recalled when presented with this emotionally salient music (El Haj, Fasotti, & Allain, 2012; Meilan Garcia et al., 2012; Palisson et al., 2015). In order to increase the likelihood that PWD were familiar with the music being presented, they were given a choice of three recordings along a weekly theme. For example, if the theme for the week was “Travel,” the choice of songs along that theme were “In My Merry Oldsmobile,” “Leaving on a Jet Plane,” or “I’ve Been Working on the Railroad.” They were then given pictures and objects with which to interact as the song played, in order to involve multiple sensory systems in the stimulation of memory recall (El Haj et al., 2012; Simmons-Stern, Budson, & Ally, 2010). The PWD and caregiver were guided in a reminiscence discussion using the discussion prompts outlined in Table 2 (Thaut & Hoemberg, 2014). The session was then closed with a final goodbye song, the therapist provided the caregiver with the caregiver education sheet from which to work until the next in-person NMT session (Supplementary Data 2), and any questions were answered.

Statistical Analysis

Demographic, clinical, and outcome characteristics of the cohort were examined using descriptive statistics. Importantly, the purpose of the current study was to assess feasibility of home-based NMT with a dementia population and generate pilot data for planning future studies; thus, we report exploratory correlations. Participant demographic, clinical, and outcome characteristics were compared according to whether they completed all NMT sessions (i.e., completed 6-week follow-up visit) or withdrew prior to completion of all NMT sessions; comparisons were also made between those who completed the 12-week follow-up visit and those who did not. Categorical variables were compared with two-way frequency tables and chi-square/Fisher’s exact association tests, and continuous variables were compared with t and Wilcoxon tests. Given considerable loss to follow-up in our cohort, logistic regression analysis was also performed and Kendall’s tau correlations with missingness were calculated for all variables that were at least ordinal to further explore potential associations between drop-out and baseline characteristics of the participants.

To evaluate for changes in our chosen outcome measures after NMT, mixed regression models with unstructured repeated measures covariance matrices were fit to estimate their means at each time point and to examine changes among the time points. Using this method, correlation between repeated measures within participants partially compensates for missing data, so long as the participant has at least one measurement, and the missingness is not “missing not at random.” All available data could be used in our mixed model, therefore, rather than being restricted to complete pairs of data, which would be required for paired t tests. Fixed effects in these models included study time points and random effects included an unstructured repeated measures correlation on the errors. The Satterthwaite method was used to calculate denominator degrees of freedom for the t and F tests from the mixed model. An omnibus F test was performed to simultaneously test for any differences among the study time points. Tukey–Kramer adjustments are available for controlling the family-wise error rate (FWER) among all pairwise comparisons among time points for a single outcome. All 18 participants, regardless of completion status, were included in the mixed models. Because of the small sample size, univariate pairwise comparisons, without adjusting for multiple comparisons, were also considered when the mixed model F test was marginally non-significant (p < .10), to further investigate any possible trends in the pilot data that could be further explored with a larger study in the future.

Significance level was set an alpha of .05. The study is limited by the small sample size, so tests may be underpowered and adjustments for multiple testing or missing data methods, such as multiple imputation, were not feasible. Statistical analyses were performed using SAS software (version 9.4. Cary, NC: SAS Institute Inc., 2014).

Results

Clinical Characteristics

Clinical characteristics of the PWD and their caregivers are presented in Table 3. The mean (SD) age of PWD participants was 76.9 (10.0) years and their average (SD) MoCA score was 14.4 (7.8) out of 30, ranging from 1 to 26. Nine participants were classified as mild-to-moderate dementia severity, and nine as moderate-to-severe. These classifications informed the specific NMT protocol that was used for their sessions with MT-BCs for the duration of the study. Specific dementia diagnoses for participants were as follows: Parkinson’s disease dementia (n = 8); Alzheimer’s disease (n = 6); vascular dementia (n = 2); dementia with Lewy bodies (n = 1); frontotemporal dementia (n = 1).

Feasibility and Acceptability

Seven PWD–caregiver dyads (38.9%) withdrew from the study before completing all NMT sessions. Reasons for withdrawal included moving into a nursing home or hospitalization (n = 3), PWD distress with sessions (n = 3), and difficulty finding time to schedule sessions (n = 1). Two dyads completed all NMT sessions, but were lost to follow-up before the 12-week follow-up visit (entered a nursing facility [n = 1], unable to contact [n = 1]). Using logistic regression, missing any follow-up visit (6 or 12 weeks) was marginally statistically non-significantly correlated with increased age and NPI scores (p = .09 and .06, respectively). Using Kendall’s tau correlations, missing any follow-up visits (6 or 12 weeks) correlated with higher neuropsychiatric symptoms burden (NPI, r = .43, p = .04) and lower caregiver-rated quality of life of the person with dementia (QOL-AD: Informant, r = −.41, p = .05). No other demographic, clinical, or outcome measure features affected the likelihood of completing the 6- or 12-week follow-up visit.

Participants who withdrew prior to completing all NMT sessions were significantly older (age, mean difference 9.4 years, t_15.05 = 2.28, p = .04) and had higher NPI scores (mean difference 17.4 points, t_16.00 = 2.85, p = .01) at baseline than those who completed all NMT sessions. Logistic regression found that a 1-year increase in participant age increased the odds of withdrawing from NMT by an estimated 13.1% (p = .03), and an increase in 1 point on the NPI increased the odds by an estimated 9.2% (p = .02). Kendall’s tau correlations found similar results: missingness at the 6-week follow-up correlated with higher NPI (r = .51, p = .01), and correlation with age was marginally statistically non-significant (r = .37, p = .07). No other demographic, clinical, or outcome measure features affected the likelihood of completing all NMT sessions.

Effectiveness

Using mixed models, there were no significant improvements in any outcome measures during the course of the study, though NPI (F_2,12 = 3.74, p = .055, Figure 1) and RSCSE (F_{2, 10.6} = 4.02, p = .051, Figure 2) approached significance for improvement. For the NPI, this is may be influenced by participants with higher baseline scores dropping out of the study, rather than true improvement with NMT for those who remained, but the correlation matrix of the mixed regression model should partially compensate since all baseline measures were included. Univariately, there were statistically significant decreases in mean NPI for the 6-week (estimate = −4.74, 95% CI: −8.43, −1.05, t_15.1 = −2.74, p = .02) and 12-week (estimate = −3.83, 95% CI: −7.41, −0.25, t_12.8 = −2.3, p = .04) visits compared to baseline (Table 3). For the RSCSE, there was an unexpected estimated decrease in caregiver-rated self-efficacy (RSCSE) at the 6-week visit after completing NMT, but then a subsequent improvement above baseline levels at the 12-week follow-up using the mixed model (Figure 2). Univariately, the improvement in RSCSE from 6 to 12 weeks was statistically significant (estimate = 135, 95% CI: 21, 249, t_10.6 = 2.63, p = .02), and the change from baseline to 12 weeks was marginally statistically non-significant (estimate = 92, 95% CI: −2, 187, t_9.01 = 2.20, p = .055) (Table 3).

Discussion

This pilot study examined the feasibility and acceptability of an in-home NMT protocol for the management of BPSD, in order to inform potential future clinical trials. While home-based NMT has been explored for motor symptoms following stroke (Street et al., 2018; Street, Magee, Odell-Miller, Bateman, & Fachner, 2015), this is the first study to apply its methods to a dementia population in the home. We have demonstrated that a NMT protocol based on MSOT, MACT, and AMMT can feasibly be performed in the home setting and may improve BPSD for community-dwelling PWD, an effect that was sustained 6 weeks after conclusion of NMT sessions. However, our findings also show that in-home NMT can be stressful and difficult to maintain for PWD, particularly when BPSD are more prominent.

Participants with more severe BPSD, as measured by the NPI, and of older age were less able to participate and more likely to drop-out prior to completing all 6 weeks of therapy sessions. Though formal qualitative research techniques were not employed for this study, members of our research team informally discussed their impressions of the feasibility of our home-based NMT protocol. Our consensus was that it could be too stressful to add a new therapeutic intervention into the home for people with more advanced dementia, many of whom were nearing the time when nursing home placement was necessary based on their care needs and caregiver burden. Furthermore, the additional burden of a weekly scheduled appointment for caregivers, even when provided within the home, should be considered. This may explain the unexpected decrease in caregiver-rated self-efficacy seen at the 6-week time point. However, it is possible that once caregivers became more comfortable with utilizing learned NMT techniques on their own and were no longer beholden to a scheduled therapy session, there was subsequent improvement in their self-efficacy.

Four PWD were hospitalized or transitioned to a nursing home during their participation in the study, indicating that we may be attempting this intervention at too late a stage of their disease. Our high drop-out rate (38.9%) suggests that initiating a new therapy when bothersome behavioral or psychological symptoms are already present (which was an inclusion criterion for this study) is not ideal, even if the therapy is provided in the home setting. Instead, introducing the home-based NMT protocol to PWD and caregivers earlier in the course of a neurodegenerative condition, even at the mild cognitive impairment stage, could be more acceptable and useful in the long term. While BPSD are less frequent at earlier stages of dementia, preparatory home-based NMT and continuation of its techniques through NMT “homework” for PWD and their caregivers could allow for more effective management of any future symptoms. In this way, PWD and their caregivers would already be familiar with the techniques of NMT, continue the methods they have learned independently, and be able to call upon these tools in times of need as the disease progresses. The availability of an effective, safe treatment option for BPSD, should they develop during the course of their disease, could serve to empower PWD and their caregivers, allaying some anxieties related to the progression of the condition. By being well prepared and taking an active role in symptomatic management through NMT, rather than relying on medications or more passive treatment modalities, PWD and their caregivers could personally effect meaningful changes in their health and well-being.

Planned next steps include a randomized controlled trial of our home-based NMT protocol at an earlier stage of mild dementia, when PWD and their caregivers can be more actively involved and receptive to treatment, to confirm our initial findings of improvements in neuropsychiatric symptoms and explore other potential benefits that were not captured with our small sample. There are also plans to explore more formalized NMT “homework” and directed instructions for PWD and their caregivers to continue on their own. Potential “hybrid” NMT protocols, with some sessions occurring in the home and some at the music therapy offices, may also be possible at earlier stages of dementia and allow for a multifaceted approach. We do believe that home-based NMT, outside of the more formal music therapy environment of a clinic or office, will help reinforce a degree of independence in the practice of NMT techniques for PWD and their caregivers, allowing for direct translation of its benefits to their daily lives. Refinement of our protocol, especially education and “homework” assignments, will be required for future studies to sustain potential benefits. Furthermore, more in-depth qualitative exploration of NMT and its impact on BPSD, through detailed interviews with PWD and caregivers, will be required to better understand if and how this treatment can be effective. For the long term, longitudinal analysis for health and quality of life outcomes for PWD and their caregivers, including decreased nursing home placement, is a future line of study in this high-need field.

The authors would like to thank the people with dementia and their caregivers who participated in this study, as well as the additional music therapists who provided treatments, Kristin Sjoberg, MA, MT-BC and Spencer Brown, MA, MT-BC. This work was supported by the ProjectSpark Foundation. The authors declare no conflicts of interest.

References